The mechanism of Bayer residue flocculation

Abstract

The aim of this study was to determine the mechanism of Bayer residue flocculation. Hematite was chosen as the test substrate as it is a common Bayer residue mineral. Batch settling tests were used to gain an understanding of the aggregation mechanism and to compare the effect of different parameters on flocculation performance. Flocculant adsorption isotherm measurements were related to changes in flocculation performance. Infrared spectroscopy was used to ascertain the configuration of adsorbed flocculant on the hematite surface.Batch settling tests showed that under strong caustic conditions hematite is naturally coagulated and that flocculation occurs via a bridging mechanism. This was confirmed by results which showed that factors which affect the bridging efficiency of the flocculant had an impact on aggregation. In particular, temperature and caustic concentration were found to greatly influence flocculation performance. This is due mainly to changes in the viscosity of the liquor, but may also be linked to the kinetics of particle-particle and particle-flocculant collisions resulting in a less efficient aggregation process. Ionic strength did not impact on performance as the flocculant was at a limiting size for synthetic liquors containing TC >/= 50 and TA >/= 10.Increasing ionic strength did not increase the adsorption density of the flocculant on hematite nor did altering the salt cation species from Na(subscript)2+ to Ca(subscript)2+. It can be concluded, therefore, that the flocculant is chemisorbed through surface complexation, since if it were electrostatically bound an increase in flocculant adsorption should have been observed with increasing ionic strength or cation charge. The surface complexation mechanism was supported by infrared results which showed that the flocculant vibrational bands were shifted on adsorption. The magnitude and direction of the shift suggests a bridging bidentate structure at pHs >/+ 11, while a monodentate structure exists at pH 7. In the presence of calcium there is also some electrostatically adsorbed flocculant at pH 7, with the calcium being in a bidentate chelating structure, but this is not observed at much higher pHs.The flocculant had an adsorption isotherm best described by a Langmuir-Freundlich expression with a monolayer coverage of ~ 164 mu g m(subscript)-2 of hematite. The adsorption density was lowered by the presence of carbonate and silicate and the action of both is thought to be due to their adsorption on active sites blocking polymer adsorption. Carbonate has an impact on flocculant adsorption at concentrations > 10 mg g(subscript)-1 while in the case of silicate ~0.2 mg g(subscript)-1 is required for the adsorption density of the flocculant to be affected. While it has been confirmed that silicate does adsorb on hematite, it was not possible to determine whether this was adsorption of a silicate species or an aluminosilicate species.X-ray photoelectron spectroscopy (XPS) showed conclusively that sodium is not involved in the adsorption of the carboxylate to the hematite surface. The lower peak shift between the backbone carbon and the carboxylate carbon suggests that the carboxylate is bonded directly to iron which has a low effective charge.Flocculant adsorption was atomistically modelled using decanoate and decandioate molecules. Modelling supported the results from XPS and infrared analysis and showed the carboxylate oxygen atoms in both organic molecules bonded directly to the surface iron atoms. Adsorption was preferred on near unhydrated surfaces with the most stable adsorption configuration being a non-symmetrical bridging bidentate structure as inferred from the infrared results.